Proxmox doesn't follow ceph.conf on external ceph cluster

[ronweister]

For the rbd pool we shouldn't have to specify --data on every single image creation. I hope proxmox by default doesn't do it as the pool being used is the default rbd pool regardless of the following setting in my ceph.conf which is followed by openstack.

[client.admin]
rbd default data pool = rbd-ec112

This should ideally work as it does with openstack. Any reason proxmox doesn't respect this rule?

 

[Alwin]

For the rbd pool we shouldn't have to specify --data on every single image creation. I hope proxmox by default doesn't do it as the pool being used is the default rbd pool regardless of the following setting in my ceph.conf which is followed by openstack.

The PVE stack (and Proxmox Support) doesn't support EC on hyper-converged setups. The tooling uses the pool 'rbd' by default and doesn't check in the ceph.conf.

[client.admin]
rbd default data pool = rbd-ec112

This should ideally work as it does with openstack. Any reason proxmox doesn't respect this rule?

For external cluster, you can put a separate config file with your storage (naming: <storageid>.conf) under '/etc/pve/priv/ceph/'. Try if it works with that.